Month: May 2017

Engineering better opioids: A podcast featuring Stanford bioengineer Christina Smolke

Obtaining compounds from nature, such as opioids from poppies or taxol from yew trees, is hard and time-consuming. So researchers, including Stanford’s Christina Smolke, PhD, are working to synthesize medically useful compounds by reengineering nature.

Smolke, a professor of bioengineering, describes her efforts to engineer yeast to make opioids on a recent episode of the “Future of Everything” radio show.

“These are compounds in nature that the opioid poppy has evolved to make. And to date, our chemists have not been able to develop efficient processes to make these compounds,” Smolke told show host Russ Altman, MD, PhD, a professor of bioengineering, of genetics, of medicine and of biomedical data science. “So we still farm this drug crop of opioid poppy to produce these molecules and the raw materials to make these molecules. And there are many limitations that come about from doing that.” These limitations include environmental and geopolitical risks, she said.

Smolke explained that she tackled this research even though many experts in the field viewed it as impossible — because it involved reengineering a complicated set of reactions and mix of enzymes that work together within the opioid poppy to build the opioid molecules. Over 10 years, her research team developed the very challenging platform technology to “prove that it could be used with any compound found in nature.”

“The final yeast strain that made the initial opioids molecules had 23 different enzymes put into it. So one of the challenges was identifying the enzymes from the opioid poppy and then moving them into yeast,” Smolke said.

But the trickiest part, she explained, was getting them to work in yeast, which is a very different organism than opioid poppies. The researchers had to modify each of the enzymes to create a yeast strain that could churn out opioid molecules.

There is more work to do though, including creating yeast that are more efficient at making the opioid compounds, as well as using the technology to make better opioids with less side effects so they are less addictive. Luckily, Smolke expects her new research projects to go more quickly now that they’ve developed the basic tools.

“We’re probably around 5 years away from molecules coming from yeast-based platforms to actually be in the medications that you’re taking,” Smolke explained. “Some of that lag is due to the engineering that we have to do to make the processes efficient enough so they can be scaled up at a commercial setting. And others are [due to] regulatory approvals.”

This is a reposting of my Scope blog story, courtesy of Stanford School of Medicine.

Mowing down cancer: A podcast featuring Stanford chemist Carolyn Bertozzi

To explain her work, Stanford chemistry professor Carolyn Bertozzi, PhD, often turns to analogies. Cancer cells, she says, are like M&Ms with a hard sugar coating. As she recently explained on the “Future of Everything” radio show, the coating’s function has remained a mystery for years, but now researchers are making real progress.

“We have come to think of these sugars as kind of a 2D barcode. The patterns are different on different cell types, and yet all of the cells of a certain type have a common pattern,” Bertozzi told show host Russ Altman, MD, Phd. “So there is a code there, but we don’t quite have the means to scan it and we don’t yet understand it.”

So what do the barcodes look like on cancer cells? Bertozzi describes them as a superposition of two barcodes — the original cell’s barcode and a new cancerous one. And the cancerous barcode looks similar for many different cancers. Researchers have found that these sugar barcodes on cancer cells can promote disease progression by turning off the immune system. “They basically tell immune cells, ‘There’s nothing to see here. Move along. I’m perfectly fine and healthy,’” Bertozzi said.

Using an analogy, she explained in the podcast that the cancer cells put on makeup to look fabulous and mesmerize the immune system, fooling it into thinking that the cells are healthy so the cancer can progress unimpeded. Her lab is developing a way to strip off this makeup.

Her team has developed a way to use enzymes to cut off the sugars, making the cells available for immune cells to target. She explained: “They were enzymes that normally play a role in digesting sugars. So what we’ve done is repurposed these enzymes so we can target them right to the surface of the cancer cell. And literally they’ll just go across the surface of the cell mowing off the sugars, like stripping off the makeup. And then the cells can be seen for what they truly are.”

Bertozzi is also involved in a company that hopes to bring this “lawn mower” technology to the clinic within the next two years, but they first need to get good preclinical data as proof-of-concept. The company is currently focused on developing new treatments for breast, lung and kidney cancers.

This is a reposting of my Scope blog story, courtesy of Stanford School of Medicine.

Atrial fibrillation more common than previously reported, study finds

Photo by BruceBlaus

Atrial fibrillation (Afib), the most common type of heart arrhythmia, affects millions of Americans. People with Afib can experience an irregular heartbeat, heart palpitations, shortness of breath, lightheadedness, fatigue and chest pain.

However, some patients with Afib have no symptoms — making it difficult to diagnose the disease early enough to overcome the increased risk of life threatening conditions such as heart failure and stroke. New research suggests this may be a bigger problem than previous thought.

“The incidence and prevalence of Afib have not been well defined as patient symptoms are not a reliable indicator of Afib,” said Javed Nasir, MD, a Stanford cardiac electrophysiology fellow. “Most Afib episodes are asymptomatic and most symptoms thought by patients to be Afib are actually not associated with the arrhythmia. Furthermore, Afib is an intermittent disease and doesn’t lend itself to robust detection with traditional intermittent monitoring modalities, such as ECG or Holter monitors.”

To see just how widespread undetected Afib may be, Nasir led a screening trial using insertable cardiac monitors (ICMs). ICMs are single-lead ECG monitoring devices, about one-third the size of an AAA battery, that are inserted under the skin of the chest. The devices can automatically detect and record Afib episodes and can remotely transmit the data to a doctor’s office.

“Recently there have been significant advances in technology and we now have very small ICMs with the ability to continuously monitor for Afib for years,” Nasir said. “We started this trial with the hopes of using this technology to identify a population with a high risk of Afib.”

The research team used ICMs to screen almost 250 elderly people with a mean age of 74 years and with no history of atrial fibrillation. They followed the patients for 18 months with monthly remote analysis of the ICM data that was reviewed by cardiologists. As recently reported in Heart Rhythm, they found that 22 percent of the participants were newly diagnosed with atrial fibrillation.

“While classically we could give a 40 year old adult a 25% chance of developing Afib in their lifetime, in our trial we nearly saw this with only 18 months of monitoring,” Nasir said.

The study also demonstrated that the majority of these newly diagnosed Afib patients were then treated with oral anticoagulants, which have been shown to significantly reduce the risk of stroke in patients with Afib detected with pulse palpation or an ECG.

Yet more research is needed, Nasir said:

“We have begun to appreciate that stroke risk varies with the amount of Afib, and the Afib found with ICM screening tends to be brief and asymptomatic. Before we recommend routine screening for Afib with ICMs, it is important to wait for the results of on-going trials that are evaluating the risks and benefits of oral anticoagulants in patients with device detected Afib. And we will need to carefully consider the costs of screening with ICMs.”

This is a reposting of my Scope blog story, courtesy of Stanford School of Medicine.

Researchers develop technology capable of real-time drug level monitoring and maintenance

Photo courtesy of Soh Lab

Doctors often struggle to choose the best dose of a drug for each patient — the dose that worked for patient A isn’t enough for patient B, or it is way too much for patient C. The response is governed by a host of factors, including genetics, age, body size, the use of other medications, the presence of diseases and the development of drug tolerances.

Now, Stanford researchers are developing new technology to help deliver an optimal, personalized drug dose. Using their drug delivery system, they were able to automatically administer chemotherapy at the desired concentration in mice, as reported today in Nature Biomedical Engineering.

“This is the first time anyone has been able to continuously control the drug levels in the body in real time. This is a novel concept with big implications because we believe we can adapt our technology to control the levels of a wide range of drugs,” said H. Tom Soh, PhD, senior author and a Stanford professor of radiology, of electrical engineering and of chemical engineering, in a recent news release.

The new technology uses three basic elements to create a closed-loop drug delivery system that continuously monitors and adjusts the infusion rate of the drug as needed.

First, a real-time biosensor measures the concentration of the drug in the bloodstream, using aptamer molecules that bind to a specific target molecule. (Aptamers are like antibodies made out of nucleotides.) When the drug of interest is present in the bloodstream, the aptamers bind to the drug, change shape and cause an electrochemical signal that is detected by an electrode. The more drug present, the more aptamers bind and the larger the detected signal.

Second, a controller with sophisticated software uses this detected signal to determine the optimal drug delivery rate to maintain the desired drug concentration. Third, a programmable infusion pump delivers the drug at the rate specified by the controller.

Although the initial results are very promising, many years of additional research will be needed before the system can be tested on humans. The team also plans to make many improvements, including miniaturizing the device. Currently their system is suitable for chemotherapy drug delivery — using a biosensor the size of a microscope slide, as shown in the photograph — but it is too large to be worn by a patient for continual use.

Still, the authors believe their system could be safely used in humans in the future. They stated in the paper that it would be especially useful for the controlled delivery of chemotherapy drugs to pediatric cancer patients, who are particularly difficult to dose correctly since their drug response varies widely with age and degree of physical development.

This is a reposting of my Scope blog story, courtesy of Stanford School of Medicine.

Stanford study provides new understanding of breast growth disorders

Photo by sasint

Breast underdevelopment at puberty is associated with a shortage of several hormones produced by the pituitary gland, a condition called combined pituitary hormone deficiency (CPHD). This disorder is caused in part by loss-of-function mutations of the GLI2 gene, but the molecular pathways of how CPHD manifests are not fully understood.

Now, researchers at Stanford University School of Medicine have discovered a new way that GLI2 impacts breast development, as recently reported in Science. Led by Philip Beachy, PhD, a Stanford professor of developmental biology and of biochemistry, the research team found that GLI2 activity helps control mammary stem cells in mice.

Stem cells are responsible for the growth, homeostasis and repair of many tissues. The behavior and survival of these stem cells depends on their local microenvironment, called a stem cell niche. During breast growth, the niche must support its associated stem cells while also responding to circulating hormones that trigger the dramatic changes of puberty.

The study showed that this stem cell niche is genetically programmed to produce the signals that control breast development in response to the hormones that regulate puberty. Using mice without a functioning GLI2 gene, the researchers found that a defective stem cell niche environment may lead to the breast growth defects seen in human CPHD. In addition, the research provides insights into a new mechanism to target when developing drugs that may help prevent breast cancer.

The authors conclude:

“Whereas prior studies implicate stem cell defects in human disease, this work shows that niche dysfunction may also cause disease, with possible relevance for human disorders and in particular the breast growth pathogenesis associated with combined pituitary hormone deficiency.”

This is a reposting of my Scope blog story, courtesy of Stanford School of Medicine.

 

Artificial intelligence could help diagnose tuberculosis in remote regions, study finds

Image courtesy of Paras Laknani

Tuberculosis is an infectious disease that kills almost two million people worldwide each year, even though the disease can be identified on a simple chest X-ray and treated with antibiotics. One major challenge is that TB-prevalent areas typically lack the radiologists needed to screen and diagnose the disease.

New artificial intelligence models may help. Researchers from the Thomas Jefferson University Hospital in Pennsylvania have developed and tested an artificial intelligence model to accurately identify tuberculosis from chest X-rays, such as the TB-positive scan shown at right.

The model could provide a cost-effective way to expand TB diagnosis and treatment in developing nations, said Paras Lakhani, MD, study co-author and TJUH radiologist, in a recent news release.

Lakhani performed the retrospective study with his colleague Baskaran Sundaram, MD, a TJUH cardiothoracic radiologist. They obtained 1007 chest X-rays of patients with and without active TB from publically available datasets. The data were split into three categories: training (685 patients), validation (172 patients) and test (150 patients).

The training dataset was used to teach two artificial intelligence models — AlexNet and GoogLeNet — to analyze the chest X-ray data and classify the patients as having TB or being healthy. These existing deep learning models had already been pre-trained with everyday nonmedical images on ImageNet. Once the models were trained, the validation dataset was used to select the best-performing model and then the test dataset was used to assess its accuracy.

The researchers got the best performance using an ensemble of AlexNet and GoogLeNet that statistically combined the probability scores for both artificial intelligence models — with a net accuracy of 96 percent.

The authors explain that the workflow of combining artificial intelligence and human diagnosis could work well in TB-prevalent regions, where an automated method could interpret most cases and only the ambiguous cases would be sent to a radiologist.

The researchers plan to further improve their artificial intelligence models with more training cases and other artificial intelligence algorithms, and then they hope to apply it in community

“The relatively high accuracy of the deep learning models is exciting,” Lakhani said in the release. “The applicability for TB is important because it’s a condition for which we have treatment options. It’s a problem that we can solve.”

This is a reposting of my Scope blog story, courtesy of Stanford School of Medicine.